Become size wise— Adopt COSMIC FFP for estimation

A simple model for functional sizing from the Common Software Measurements International Consortium (COSMIC) that overcomes the limitations of all other traditional methods

Annual CHAOS studies over the last decade reveal that only 16-18 percent of the projects get delivered within the allowed cost, quality and time schedules. For the 50 percent of projects that manage completion, cost and time overruns on an average stand at 200 percent. The software economy, which now holds a significant part of the overall economy, adversely affects the global economy through this colossal wastage. One of the root causes of the failure is ineffective project planning and estimation.

Function Points User Group (FPUG) has of course made valiant efforts to produce guidance on how to interpret Albrecht's original five component-types in terms of modern development methods. But inevitably, if the developer is working in terms of concepts such as Objects or GUIs or Use Cases or APIs or whatever, the translation from these concepts back to the three-decade-old concepts of 'Elementary Inputs' or 'Internal Logical Files' will seem like a distraction to the main purpose.

Under the umbrella of Common Software Measurements International Consortium (COSMIC), a model was brought out to introduce a new size estimation method that can overcome the limitations of all other methods; thus was born the new COSMIC FFP method.

COSMIC FFP model

The major components of the cosmic model are functional processes and data movements in terms of data groups. The COSMIC-FFP model for functional

sizing is extremely simple. Functional User Requirements are decomposed into 'Functional Processes', which in turn are decomposed into 'Functional Sub-Processes'. A Functional Sub-Process is a type of Data Movement, namely an Entry, an Exit, a Read or a Write, each assumed to have associated data manipulation.

Steps involved in Size Estimation using COSMIC FFP method are:

1. Identification of the entities / data groups in the software problem.

2 Breakdown of the requirements into functional user requirements.

3. Identification of the number of data movements in terms of entities / data groups for each functional sub-processes.

4. Calculation of Total Cfsu (Cosmic Units) in terms of aggregation of all data movements.

COSMIC FFP has two Measurement View Points i.e., 'End User View Point’ and ‘Developer View Point.’ FPA was helpful in counting size with End User View Point. Application development involves developing many building brick components, and many components that communicate among themselves in a peer-to-peer model. Projects end up developing additional functionality to meet end-user functionality, which usually does not get accounted for. This size is very critical when it comes to estimating efforts for development and project scheduling. The size delivered to customers is sometimes different from the size that he was looking for. COSMIC FFP, by bringing Layers and Peer components into its scope of estimation, provides a 'Developer View Point' of size.

Complex applications are no more just a business application or real-time application. They are a hybrid of various domains resulting in solutions to a customer's problems. When more hardware and devices get integrated with the everyday lives of people, applications become MIS-type business applications, but with real-time characteristics. Mobile application is real-time when it comes to managing a call, and a database application with respect to call logging and billing.

It is possible to size such an application using COSMIC FFP.

COSMIC FFP has been announced as International Standard - ISO 19761:2003. The method merely states the obvious: You have to know your application right from the beginning to a fair amount of detail. The software industry is much used to the fact that this is not possible, whereas COSMIC FFP is a method which by its inherent nature leads you to asking a lot of questions and getting most of the requirements clear at the very beginning of the project. Thus, the method is a welcome change to the world of MIS software and a good introduction to the world of real-time software.

The author is a member of the International Advisory Council of the Common Software Measurements International Consortium (COSMIC). E-mail: jayakumar@amitysoft.com